Theodore Vasiloudis created FLINK-2202:
------------------------------------------
Summary: Calling distinct() requires tuple input
Key: FLINK-2202
URL: https://issues.apache.org/jira/browse/FLINK-2202
Project: Flink
Issue Type: Improvement
Components: Core, Scala API
Reporter: Theodore Vasiloudis
Priority: Minor
Currently to call distinct on a DataSet the elements must be placed in a tuple.
This creates the need to write code like the following:
{code}
val doubleDS: DataSet[Double] = ...
val uniqueDS = doubleDS.map( el => Tuple1(el)).distinct().map(t => t._1)
{code}
which looks quite unnecessary. Ideally we would like to just have to write:
{code}
val uniqueDS = doubleDS.distinct()
{code}
which should be possible as long as there exists an implicit {{Ordering\[T\]}}
for a {{DataSet\[T\]}}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
